US8351084B2 - Vector error diffusion with edge control - Google Patents
Vector error diffusion with edge control Download PDFInfo
- Publication number
- US8351084B2 US8351084B2 US12/577,798 US57779809A US8351084B2 US 8351084 B2 US8351084 B2 US 8351084B2 US 57779809 A US57779809 A US 57779809A US 8351084 B2 US8351084 B2 US 8351084B2
- Authority
- US
- United States
- Prior art keywords
- pixel values
- marking
- contone
- marking decision
- error
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related, expires
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/40—Picture signal circuits
- H04N1/409—Edge or detail enhancement; Noise or error suppression
- H04N1/4092—Edge or detail enhancement
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/46—Colour picture communication systems
- H04N1/52—Circuits or arrangements for halftone screening
Definitions
- the presently disclosed embodiments are directed toward techniques for quantizing or transforming continuous tone or contone image data for rendering on reduced quantization rendering devices such as liquid crystal displays and binary or high-addressable printers.
- Embodiments will be described with regard to subtractive colorant (e.g., cyan, yellow and magenta) based rendering devices (e.g., printers).
- Embodiments are applicable to other colorant spaces, including, but not limited to the red, green and blue color spaces of display devices and the additional colorant dimension color spaces of high fidelity (HI-FI) and Photo Tone printing devices.
- Image quantization techniques include halftoning via screening and error diffusion techniques. These techniques determine or select marking decisions. Marking decisions can be binary (on or off, mark or no mark) or be associated with high addressability pixels associated with additional quantization levels. However, marking decisions are associated with fewer quantization levels that are contone pixel values. Accordingly, selection of a set of marking decisions for an image is said to result in a reduced quantization version of that image.
- Each class of technique has advantages and disadvantages as compared to the other class.
- halftoning can lead to visual artifacts. For instance, halftone cell repetitions often fall within a range of frequencies that are visually detectable. Error diffusion often results in marks being placed at a frequency and in a non-repeating manner that is less visible. However, error diffusion techniques can be associated with other objectionable phenomena.
- Eshbach describes a method for edge modification (i.e., enhancement or attenuation) that is applicable to scalar error diffusion techniques.
- scalar refers to quantization techniques which consider only a single value from a given pixel when making a marking decision regarding that value of that pixel. For example, scalar error diffusion techniques are applied to black and white or monochrome contone images to make marking decisions regarding a single colorant. Additionally, scalar error diffusion techniques can be applied separately to each separation of a color image.
- a scalar form of error diffusion makes marking decisions regarding the cyan colorant without considering the values of the target pixel associated with the other colorants (e.g., magenta, yellow and optionally black).
- a marking decision is made with regard to the magenta colorant without considering the values of the pixel related to the cyan, yellow and black colorants.
- Eshbach's technique varies a marking decision threshold applied to an error modified pixel value according to the contone value of the related original contone image pixel, thereby providing a means for influencing the location of lightness or darkness in the vicinity of an edge.
- scalar error diffusion techniques do not consider or compensate for the effects of interactions between dot patterns of different color separations. Accordingly, scalar techniques can suffer from color errors and displeasing moiré and other patterning artifacts.
- contone values regarding a first colorant can have a bearing or be a factor in a marking decision associated with a second colorant (e.g., magenta). Indeed, all the values of a particular pixel may be considered when making a marking decision regarding a particular colorant. That is, in vector error diffusion, colors are treated as points in a multidimensional color space with the colors printable or displayable located at discrete reference locations within that space. When a continuous tone or contone color is to be displayed, for example, the closest displayable color can be selected for display, with color error being calculated as a vector in the color space and diffused to neighboring pixels.
- a first colorant e.g., cyan
- a second colorant e.g., magenta
- vector error diffusion techniques assign marking decisions to various locations in their respective color spaces and in so doing, move away from the concept of thresholding. Accordingly, the method of Eshbach is not applicable to vector error diffusion and vector error diffusion techniques have been without a mechanism for controlling or modifying (i.e., enhancing or attenuating) edges.
- a method for transforming a color contone description of an image into a reduced quantization version of the image according to a desired edge modification can include combining contone pixel values of a target pixel with contone error values, thereby generating a set of error modified pixel values and selecting marking decisions based on the set of error modified pixel values and at least one function of at least one of the contone pixel values of the target pixel, wherein a color error modified pixel value associated with at least the first colorant is a factor in selecting a marking decision regarding at least the second colorant, wherein marking decisions for a given pixel may sacrifice color accuracy in favor of placing a mark for a given set of error modified pixel values and wherein the at least one function biases the marking decision associated with a respective error modified pixel value toward making a mark as a respective value of the received contone pixel values increases in darkness when the desired edge modification is an edge enhancement and wherein the at least one function biases the marking decision associated with a respective error modified pixel value away from making a mark
- the method can also include determining color error associated with the selected marking decision, storing the determined color error for use in processing a subsequent pixel of the contone image and at least one of storing or rendering the selected marking decisions. Repeating this process, whereby each pixel in an image or selected portion of an image is selected and processed as the target pixel results in transforming the image into a reduced quantization version of the image. For instance, the process can be used to transform the image into a binary or high addressable quantization version of the image.
- the described biasing provides a mechanism for increasing or decreasing the responsiveness of a vector error diffusion process to edges or changes in color as processing moves along a processing path through the image.
- An image processing system that is operative to transform a color contone description of an image into a reduced quantization version of the image according to a desired edge modification can include a pixel value combiner, a marking decision selector and color error determiner.
- the pixel value combiner can be operative to receive contone pixel values of target pixels of the image which call for colorant densities of a plurality of colorants of an associated image rendering device, the plurality of colorants including a first colorant associated with at least a first pixel value of the target pixel and a second colorant associated with at least a second pixel value of the target pixel, to receive contone color error values associated with marking decisions made in regard to a previously processed contone pixel of the image or initial contone error values and to combining the contone pixel values with the contone error values, thereby generating a set of error modified pixel values.
- the marking decision selector can be operative to select marking decisions based on the sets of error modified pixel values and at least one function of at least one of the received contone pixel values of the target pixels, wherein a color error modified pixel value of a target pixel associated with at least the first colorant is a factor in selecting a marking decision regarding at least the second colorant of the target pixel, wherein marking decisions for a given pixel may sacrifice color accuracy in favor of placing a mark for a given set of error modified pixel values and wherein the at least one function biases the marking decision associated with a respective error modified pixel value toward making a mark as a respective value of the received contone pixel values increases in darkness when the desired edge modification is an edge enhancement and wherein the at least one function biases the marking decision associated with a respective error modified pixel value away from making a mark as a respective value of the received contone pixel values decreases in darkness when the desired edge modification is an edge attenuation.
- the color error determiner can be operative to determine color error associated with each selected marking decision and make contone color error values describing the determined color error available for the processing of at least one subsequently selected target pixel.
- FIG. 1 is a flow diagram outlining a method for transforming a contone description of an image into a quantized version of the image.
- FIG. 2 is a flow diagram outlining aspects of an embodiment of the method of FIG. 1 .
- FIG. 3 depicts aspects of illustrative marking decision color space definitions.
- FIG. 4 illustrates an equivalency between a relocation of an error diffusion threshold and a relocation of color space reference points.
- FIG. 5 illustrates a relocation of reference points in a three dimensional marking decision color space.
- FIG. 6 illustrates an aspect of a relocation of reference point in a marking decision color space.
- FIGS. 7 and 8 outline aspects of an embodiment of the method of FIG. 1 .
- FIGS. 9 and 10 illustrate an equivalency of changing locations of reference points in a marking decision color space and offsetting values of a pixel being processed according to the marking decision color space.
- FIG. 11 is a block diagram depicting an embodiment of a system, such as an image processing device, that is operative to perform embodiments of the method of FIG. 1 .
- a method 110 that is operative to transform a color contone description of an image into a reduced quantization version of the image according to a desired edge modification can include receiving 114 contone pixel values of a target color pixel, receiving 118 color contone error values, combining 122 the received target pixel contone values with the error contone values to generate error modified pixel values and selecting 126 a marking decision based on both the error modified pixel values and at least one function of at least one value of the received target pixel contone values, wherein an error modified pixel value regarding at least the first colorant is a factor in selecting a marking decision regarding at least a second colorant.
- the selection 126 of a marking decision is made according to a variation of vector error diffusion wherein a pixel value related, for example, to a cyan colorant can have a bearing on or be a factor in a marking decision related to, for example, a magenta colorant.
- the at least one function of at least one value of the received target pixel provides a mechanism for adjusting the vector error diffusion technique to provide a mechanism for modifying (enhancing or attenuating) the appearance of edges in the image.
- the method 110 further includes determining 130 a color contone error associated with the selected 126 marking decision, storing 134 the determined contone color error for use in processing one or more subsequently processed pixels and storing or rendering 138 the selected marking decision.
- This process ( 114 - 138 ) can be repeated for each pixel in the image being processed. As each pixel is processed, at least one value of the respective received 114 target pixel separately and individually influencing the marking decision selection 126 associated with that respective target pixel according to the received 114 contone pixel values of that target pixel.
- selecting 126 a marking decision based on both the error modified pixel values and at least one function of at least one value of the received target pixel can include defining 214 a marking decision color space relating marking decisions to contone values regarding a plurality of colorants, the color space being defined, at least in part, by respective reference points, the respective reference points being associated with respective extremes of ranges of available contone values and therefore with respective marking decisions having zero error.
- a marking decision color space associates respective subsets of sets of contone pixel values with the marking decisions of respective ones of the respective reference points.
- the defined 214 marking decision color space is associated with a reference set of contone pixel values.
- the marking decision color space is used as defined 214 to select 126 a set of marking decisions for a set of error modified pixel values when a received 114 set of contone pixel values of a target pixel corresponds to the reference set of contone pixel values.
- the reference set of contone pixel values may describe a white or completely unmarked pixel or image portion.
- Selecting 126 the marking decision based on both the error modified pixel values and at least one function of at least one value of the received target pixel can further include changing 218 the location of at least one reference point in the marking decision color space according to a function of a difference between the reference set of contone pixel values (e.g., white) and the contone pixel values of the target pixel.
- a movement of at least one reference point changes the distance or relative position in the marking decision color space between points associated with at least one of the respective subsets of contone pixel values and the at least one reference point and therefore may change at least one marking decision associated with at least one set of contone pixel values.
- Such a change 218 or relocation of a reference point and marking decision association generates a warped marking decision color space relative to the defined 214 marking decision color space and results in an edge modification (enhancement or attenuation) when the warped space is used in a VED process.
- selecting 126 a marking decision based on both the error modified pixel values and at least one function of at least one value of the received target pixel further includes identifying 222 a point in the warped marking decision space corresponding to the error modified pixel values and selecting 226 the marking decision associated with the identified point according to the warped marking decision space.
- the marking decision color space can include defining a marking decision color space, at least in part, by the location of respective reference points associated with extremes of ranges of available contone values and therefore with respective marking decisions having zero error.
- contone values typically range from 0 to 255. The value 0 is associated with not placing a mark and the value 255 is associated with placing a mark.
- R red
- B blue
- Some reference points are associated with a zero error marking decision to place a mark for one and only one colorant.
- C cyan
- the cyan reference point is associated with placing a mark for one and only one colorant.
- the marking decision color space 310 is illustrative only and that other color spaces can be defined for other applications.
- the colorants are light emitting diodes or phosphors of a display device
- reference points representing red, green and blue might be associated with decisions to place a mark using one and only one colorant (e.g., a red, green or blue phosphor or LED)
- Cyan, magenta and yellow reference points might be associated with marking decisions to place marks using combinations of two colorants.
- a yellow reference point might be associated with marking decisions to place marks using a red and green colorant and not place a mark using a blue colorant.
- color spaces of a higher dimensionality are also possible.
- a color space can be defined 214 to associate marking decisions with various sets of contone values in color spaces involving the additional colorants available in Hi-Fi rendering devices.
- a marking decision color space can be defined 214 in just two dimensions.
- vector error diffusion techniques assign marking decisions to contone pixel values, or sets thereof, in a manner that biases marking decisions toward placing marks for a given pixel using fewer colorants than would otherwise be used.
- VED vector error diffusion
- a marking decision color space e.g., 310
- a marking decision color space is segmented or partitioned in a manner that biases marking decisions toward decisions to use or place marks for fewer colorants.
- Such a marking decision is too light in the cyan separation. Accordingly, error is diffused to one or more neighboring pixels in a manner that darkens or increases an associated cyan pixel value of the one or more neighboring pixels. Accordingly, the chances are increased that a marking decision will be made to place a mark using a cyan colorant when the one or more neighboring pixels (or error diffused versions thereof) are processed. Moreover, since the marking decision color space (e.g., 310 ) is defined in a manner biased toward placing marks using a reduced number of colorants, as the neighboring pixels are processed, additional errors in darkness will be associated with the processing of the neighboring pixel and will be propagate to pixels neighboring those pixels. Accordingly, as will be discussed in greater detail below, VED increases the spatial frequency of marks and reduces the visibility of imaging artifacts.
- the space 310 is segmented or partitioned by planes. Each plane intersects a surface associated with a contone value of zero for one colorant and dividing that surface into portions associated with the other colorants. For example, a first plane 346 intersects a surface associated with a contone value of zero for a yellow colorant and divides the surface between the magenta and cyan colorants. A second plane 350 intersects a surface associated with a contone value of zero for the magenta colorant and divides the surface between the cyan and yellow colorants.
- a third plane 354 intersects a surface associated with a contone value of zero for the cyan colorant and divides the surface between the magenta and yellow colorant.
- the point 342 discussed above is closer to the reference point 334 associated with placing a mark using the magenta colorant and only the magenta colorant than are any points on the planes 346 , 350 , 354 located as measured along the direction of a vector from that reference point 334 to the point 342 . Therefore, according to the definition of the illustrative marking decision color space 310 , the point 342 is associated with the marking decision of the reference point 334 .
- a marking decision color space definition can associate respective sets of contone pixel values with a marking decision associated with a respective reference point based on respective functions of distance in the color space of points identified by the sets of contone pixel values to respective ones of the reference points.
- This segmentation is illustrative only. Other marking decision color space definitions including other segmentations are contemplated.
- the reference points ( 314 , 318 , 322 , 326 , 330 , 334 , 338 ) are associated with contone values at the extremes of available ranges and also with zero error marking decisions. That is, a contone value of 255 in an 8 bit system indicates a desire for a 100% colorant density and a mark placed for a given pixel (in a binary rendering device) represents a 100% colorant density. A contone value of zero represents a desire for a 0% colorant density. A marking decision not to place a mark exactly corresponds to a 0% colorant density. In contrast, a contone value of, for example, 64 represents a desired colorant density of about 25%.
- a binary rendering device can only place a mark or not place a mark in response to such a contone value.
- a decision not to place a mark means that the associated portion of the image will be rendered to be 25% too light.
- a decision to place a mark in response to a contone value of 64 would result in a corresponding portion of a rendered image being 75% too dark.
- compensating offsets are diffused to pixels that neighbor pixels associated with such marking decision errors.
- a pixel is rendered too darkly, contone values associated with neighboring pixels are reduced by a corresponding amount thereby lightening the neighborhood of the pixel to compensate for its excessive darkness.
- contone values associated with neighboring pixels are increased thereby increasing the chances that marks will be placed for those pixels, thereby darkening the vicinity of the original pixel in a manner that compensates for its excessive lightness. Pixels that receive such compensating offset values are referred to herein as error diffused pixels.
- contone values associated with error diffused pixels can extend far beyond the available contone value range.
- contone pixel values can be lightened below zero or darkened above 255.
- the illustrative marking decision color space 310 can extend to infinity in all directions but is defined by the reference points (e.g., 314 - 338 ) and the surfaces and segmentations associated therewith.
- selecting 126 a marking decision based on both the error modified pixel values and at least one function of at least one value of the received target pixel contone values can include defining 214 a marking decision color space (e.g., 310 ) wherein the marking decision color space and associated marking decisions are associated with a reference set of contone pixel values.
- the location of at least one reference point (e.g., 314 - 338 ) in the marking decision color space (e.g., 310 ) is changed 218 according to a function of a difference between the reference set of contone pixel values and the contone pixel values of a target pixel.
- a reference color such as, for example, white. That is, the illustrative marking decision color space 310 can be defined 214 to be the marking decision color space to use in selecting 126 a marking decision for an error diffused pixel when the associated received 114 contone pixel values of
- the warping and the associated change in marking decision mappings or assignments can be used to modify (e.g., enhance or attenuate) edge as part of a process for transforming a color contone description of an image into a reduced quantization version of the image using vector error diffusion (VED).
- VED vector error diffusion
- points in the illustrative marking decision color space 310 to the right of, for example, the surface bounded by the C 330 , G 326 , K 314 and B 324 reference points are associated with marking decisions to place marks for more colorants than are points to the left of that surface. This observation is all the more true for points on either side of planes parallel to referenced CGKB surface and inside the boundary of the illustrative cube.
- warping or transforming that decision space ( 310 ) by moving that saturated surface (saturated because it is associated with a saturated contone value (e.g., 255) for the cyan colorant at all points on the surface) toward its opposing secondary surface (i.e., the surface defined by the W 318 , M 334 , R 322 and Y 338 reference points) tends to increase the chances that decisions will be reached to place marks for a given set of pixel values. Accordingly, warping a marking decision color space in the above-described manner as image data transitions from a region of white to a region including cyan increases the chances that a decision to place a cyan mark will be made. Therefore, the transition or edge including cyan will be enhanced.
- warping the marking decision color space in this direction reduces the chance that a decision is reached for a given pixel to place a cyan mark. Accordingly, warping a marking decision color space in this manner tends to attenuate an edge including cyan. Similar warpings or combinations thereof can be applied to the other saturated surfaces in regard to other primary colorants.
- the warpings and associated marking decision selections 126 can be based on one or more functions of at least one of at least one received 114 contone pixel value of a target pixel.
- warping functions can be based on one or more functions of at least one of at least one received 114 contone pixel value of a target pixel.
- warping functions associated with movements along the cyan, magenta and yellow axes of the illustrative marking decision space 310 generally as f c (•), f m (•), and f y (•), respectively.
- the • notation indicates some argument for the function.
- Such a warping function would allow an amount of magenta in a received image pixel to be a factor in the warping of the image space along the cyan axis.
- Positive functions will be those that enhance edges while negative functions will be those that attenuate edges.
- positive functions are those where f i (•) shifts the saturated plane associated with colorant i toward its corresponding secondary plane.
- Negative functions are those where f i (•) shifts the saturated plane away from its corresponding secondary plane.
- f i (•) can take the form of algebraic, exponential or trigonometric functions.
- selected warping functions can change the location of one or more reference points beyond the range of the input pixel values (e.g., more than 255 in our illustrative 8 bit implementation). Such situations can be addressed with a more complicated warping as will be explained below.
- each function i.e., f c , f m , and f y
- the functions can be thought of as operating by taking each value on a decision space (e.g., 310 ) axis and mapping those values to new values for the axis. Thought of this way, the partitions in the space maintain their shape, but the axes values that index into those partitions are changed.
- Another equivalent form is to “warp” error modified pixel values in a reciprocal manner. The following description continues the description of the warping or changing of respective locations of respective reference points. Subsequently, an indication of how the above-mentioned and other equivalent forms can be implemented will also be provided.
- a threshold 410 is set to be between two extremes of the available colorant range 414 , 418 .
- the threshold might be set at 127.5. If an error modified pixel value (e.g., k e ) is less than the threshold, then a marking decision is made to place no mark (e.g., render a white portion of an image). If the error modified pixel value is not less than the threshold, then a decision is made to place a mark (e.g., render the associated portion of the image as black).
- an equivalent result is achieved by determining whether the error modified pixel value is closer to one extreme (e.g., 414 ) or the other (e.g., 418 ). That is, if the difference between the low contone value or white reference point contone value and the error modified pixel value is less than the difference between the error modified pixel value and the other extreme 418 (e.g., the contone value representing 100% black), then a decision is made to place no mark. If that condition is not met, then a decision is made to place a mark.
- one extreme e.g., 414
- the other extreme 418 e.g., the contone value representing 100% black
- a marking decision space warping effect can be achieved in single colorant applications by warping the position of the threshold to a warped threshold position 422 .
- the threshold position is warped according to a warping function f k (k i ), where k i is a value from an input pixel associated with the position of the error modified pixel value.
- a decision to place no mark would be made if the error modified pixel value is less than the warped threshold value and a decision to place a mark would be made otherwise.
- VED vector error diffusion
- a comparison of distances expression that achieves equivalent results to the warping of the threshold value is achieved by warping or changing 218 the position of a reference point (e.g., the black K 418 reference point).
- the threshold was moved closer to the white reference point by an amount equal to f k (k i ).
- f k the position of the black reference point
- the warped position 430 of the black reference point becomes its original position minus the above-described offset or 255 ⁇ 2f k (k 0 ).
- a marking decision color space e.g. 310
- Case (a) covers instances wherein the warping or location change (e.g., 2f(•)) is less than the original separation along the respective axis of movement (e.g., less than 255 contone value counts in the illustrated embodiment).
- Case (b) addresses those instances wherein the warping or location change 218 is greater than that original separation.
- the warping or location changing 218 is done on the basis of axes of the defined 214 marking decision color space. That is, reference points associated with primary surfaces are moved the same amount in the direction of a given primary axis (e.g., the C, M or Y axis). For example, the cyan, green, blue and black (CBGK 330 , 324 , 326 , 314 ) reference points move the same amount along the cyan direction (toward or away from its opposing secondary plane (e.g., WMYR 318 , 334 , 338 , 322 )).
- CBGK 330 , 324 , 326 , 314 cyan, green, blue and black
- the plane defined by the green, yellow, red and black reference points (GYRK 326 , 338 , 322 , 314 ) is moved as a unit in the direction of the yellow axis either toward or away from its opposing secondary plane (CWMB 330 , 318 , 334 , 324 ).
- the location of at least one reference point is changed 218 by translating one or more planes of the illustrative cube-shaped region of the defined 114 marking decision color space (e.g., 310 ).
- the saturated cyan plane (CBGK) and associated reference points are translated towards or away from the red plane (RMYW) by an amount indicated by 2f c (c 0 ).
- the saturated magenta plane (MRBK) and associated reference points are translated towards or away from the green plane (CGYW) by an amount indicated by 2f m (m 0 ) and the saturated yellow plane (YRGK) and associated reference points are translated towards or away from the blue plane (BCMW) an amount indicated by 2f y (y 0 ).
- the warping value for one or more of the axes is equal to the maximum contone value (e.g., 255)
- the position of the secondary plane and the warped position of its corresponding saturated plane coincide or overlap one another.
- the contone value associated with 100% colorant density e.g., 255
- the surfaces of the marking decision color space are collapsed to a single point in the marking decision color space at, for example, (0,0,0). The effect this warping or position changing 218 has on marking decision associations will be further described below.
- the associated saturated plane (CBGK, MRBK and/or YRGK) takes the place of the corresponding secondary plane (e.g., RMYW, GCYW or BCMW, respectively) and the associated secondary plane is translated away from its previous position by an amount equal to 2f(•) minus the contone value associated with 100% colorant density (e.g., 255) in the same direction that the saturated plane was moved in order to replace the secondary plane.
- the saturated cyan plane (CBGK) is translated to the position of the red plane (i.e., RMYW). That is, the new coordinates for C (e.g., 630 ) are (0,0,0); the new coordinates for B (e.g., 624 ) are (0,255,0); the new coordinates for G (e.g., 626 ) are (0,0,255); the new coordinates for K (e.g., 614 ) are (0,255,255).
- the Red plane moves 2f c (c 0 ) ⁇ 255 away from the warped CBGK plane in the Cyan direction.
- the new coordinates for W e.g., 618
- the new coordinates for M e.g., 634
- the new coordinates for Y e.g., 638
- the new coordinates for R e.g., 622
- the new coordinates for R are (255 ⁇ 2f c (c 0 ),255,255).
- Identifying 222 a point in the warped marking decision space corresponding to the error modified pixel values can include using axes associated with the space to identify a point in relation to the reference points and segmentation boundaries of the space. Since the point being identified 222 corresponds to error modified pixel values, the locating indices can have values that extend beyond the available range of available contone values (e.g., below 0 and above 255). Accordingly, the identified 222 point can be inside or outside boundaries defined by the reference points. Furthermore, since in certain applications, such as that illustrated in FIG.
- the reference points are moved closer together than they are in the reference definition (e.g., 310 ), even error modified pixel values that are within the available range of contone values (e.g., 0-255) can be associated with points that are outside the boundaries defined by the changed 218 reference point locations.
- identifying 222 the point in the warped marking decision space corresponding to the error modified pixel values can include identifying a point inside (e.g., 640 , 642 , 646 ) or outside (e.g., 650 , 654 ) boundaries defined by the location of reference points (e.g., 318 , 514 , 522 , 524 , 526 , 530 , 534 , 538 ) of the warped marking decision color space (e.g., 510 ).
- reference points e.g., 318 , 514 , 522 , 524 , 526 , 530 , 534 , 538
- Selecting 226 the marking decision associated with the identified point according to the warped marking decision space can include determining a segment of the warped space that the identified 222 point is included within. For example, a first point 640 may be included in the segment associated with placing a mark using only a magenta colorant. A second point 642 may be associated with a segment corresponding to a marking decision to place a mark using only a yellow colorant. A third point 646 may be associated with a segment of the marking decision color space corresponding to a marking decision to place a mark using only a cyan colorant. Illustrative points 650 and 654 , even though they are outside the volume defined by the reference points may also be, for example, included in a segment of the marking decision color space associated with respective marking decisions.
- defining 214 a marking decision color space can include defining region borders around selected reference points.
- region borders can be defined 714 around each reference point that is not associated with a marking decision to place a mark for one and only one colorant.
- region borders can encompass points in the marking decision color space associated with the marking decision of the respective reference point.
- respective region borders 650 - 658 are defined 714 around white 318 , red 322 , blue 324 , green 326 , and black 314 reference points of the illustrative marking decision color space 310 .
- the border regions 650 - 658 are depicted as equal sized spheres. However, the regions need not be of equal size.
- the region borders do not need to be the same shape.
- the region borders can be any regular or irregular shape and may or may not extend into or out of the region bounded by the positions of the reference points of the marking decision color space (e.g., the cube defined by the black, white, red, blue, green, cyan, magenta, yellow reference points 314 - 338 ).
- the region borders 650 - 658 are defined 714 around reference points that are not associated with a marking decision to place a mark for one and only one colorant.
- the white W 318 reference point is associated with a decision to place no marks and the black K 314 reference point is associated with the decision to place marks using three colorants.
- the red, green and blue reference points 322 - 326 are each associated with decisions to place marks using two colorants.
- the region borders are defined around each reference point that is not associated with a marking decision to place a mark for one and only one colorant.
- other marking decision color space definitions are possible.
- region borders could be defined around the reference points that are associated with marking decisions to place a mark for one and only one colorant (e.g., cyan, magenta and yellow reference points ( 330 - 338 )).
- the defined 714 region borders can serve to define regions of space external to the region bounded by the reference points (e.g., 314 - 338 ). For instance, as seen most clearly with regard to region border 650 about the white reference point 318 , a region 718 external to the region bounded by the reference point positions may also be associated with the marking decision of its respective reference point (e.g., white 318 ; no marks).
- the region is bounded by planes (e.g., 720 - 724 ) that are tangent to the region border (e.g., 650 ).
- the planes are perpendicular to respective axes of the marking decision color space 310 .
- a first plane 720 that is perpendicular to the yellow axis (which extends from, for example, the white reference point 318 to the yellow reference point 338 ).
- a second plane 722 is perpendicular to the magenta axis which extends, for example, from the white reference point 318 to the magenta reference point 334 .
- a third plane 724 is perpendicular to the cyan axis, which extends, for example, from the white reference point 318 to the cyan reference point 330 . Additional regions 726 , 728 bounded by planes 730 - 734 and 736 - 740 , respectively, that are tangent to border regions 650 , 658 about the green 326 and black 314 reference points, respectively.
- First planes 730 , 736 are perpendicular to the magenta axis.
- Second planes 732 , 738 are perpendicular to the cyan axis and third 734 and 740 are perpendicular to the yellow axis.
- Similar regions are associated with the red and blue 322 , 324 reference points and are defined by respective region boundaries 652 , 654 . However, for reasons of clarity, they are not depicted in FIG. 3 .
- those region borders can be changed 744 as the locations of their respective reference points are changed 218 .
- the region borders e.g., 650 - 658
- Such region border changes 744 generate respective warped region borders.
- a compression factor based on the function of the difference between the respective at least one value of the reference set of pixel values and the at least one value of the received contone pixel values can be calculated. For instance, where the warping or translation value (2f(•)) is less than or equal to the contone value associated with 100% area coverage (e.g., 255 ), the compression factor can be calculated to be the difference between that contone value and the warping value divided by that contone value (e.g., 255 ⁇ 2f(•)/255).
- the border regions can change their shapes. For instance, in the illustrated spherical example, the spheres may become ellipsoids. In extreme cases, one or more dimensions may be reduced to zero changing the spheres to ellipses, lines or single points.
- the warping value for one or more axes is equal to the original length of the axis or contone value associated with 100% area coverage (e.g., 255)
- the secondary plane and its corresponding saturated plane overlap each other, and the width of the ellipsoids in those directions or dimensions is equal to zero. Therefore, the ellipsoids become either a 2D ellipse (e.g., when one warping value is 255), a straight line (e.g., when two warping values are 255), or a single point in space (e.g., when all three warping values are 255).
- the general equation for the region border is adapted by dropping the term corresponding to the direction in which the warping value is 255.
- the compression factor is used to change spherical region borders as described above, and is therefore used only in terms that are squared, this rearrangement is not absolutely necessary.
- identifying 222 a point or region in the warped marking decision space corresponding to the error modified pixel values can be achieved according to a series of tests or stages.
- a first test or stage can include determining 748 whether the values of the error modified pixel values are associated with a point that falls within the boundary of one of the respective warped region borders (e.g., warped or changed 744 version of 650 - 658 ) and if it is determined to be so located, identifying 752 the respective warped region within which the error modified pixel values fall as the identified 222 point.
- the respective warped region borders e.g., warped or changed 744 version of 650 - 658
- a determination 756 can be made as to whether the error modified pixel values are associated with a point external to a region bounded by the warped or changed 218 marking decision space position of the reference points (e.g., 318 , 514 , 522 , 524 , 526 , 530 , 534 , 538 ) and within a region bounded by planes that are tangent to the warped region borders and that extend in a direction perpendicular to the direction of respective axes.
- the reference points e.g., 318 , 514 , 522 , 524 , 526 , 530 , 534 , 538
- identifying 222 the point in the warped marking decision space corresponding to the error modified pixel values can include identifying 814 the region (e.g., warped versions of, for example, 718 , 726 , 728 ) external to the region bounded by the warped marking decision space position of the reference points and within the region bounded by planes tangent to the warped region borders as the identified 222 point.
- respective distances can be determined 818 to respective reference point positions of the warped marking decision color space.
- the determination 818 can be of distances to respective reference point positions of the warped marking decision color space that are associated with a marking decision to place a mark for one and only one colorant from the position associated with values of the error modified pixel values.
- identifying 222 the point in the warped marking decision space corresponding to the error modified pixel values can include identifying 822 the region associated with the respective reference point position of the warped marking decision color space that is associated with a marking decision to place a mark for one and only one colorant that is closest to the position associated with the values of the error modified pixel values as the identified point.
- Various proximity or distance measures can be used to make each of these determinations (e.g., 748 , 756 , 818 ).
- determining 748 whether the values of the error modified pixel values are associated with a point that falls within the boundary of one of the respective warped regions can be achieved through the use of the ellipsoid equation introduced above.
- the validity of the following inequality indicates whether the values of the error diffused pixel (c e ,m e ,y e ) is within the warped or changed 714 region border, where r is the original radius of the region border as originally defined (e.g., 310 , 714 ) and k c , k m and k y are respective compression factors as discussed above.
- a determination 748 is made that the error modified pixel values (c e ,m e ,y e ) are associated with a point within a particular region boundary (i.e., the expression above is determined to be true) then that region is identified 752 to be the identified 222 point and the marking decision associated therewith can be selected 226 , 126 as the marking decision to be used to render the target pixel.
- the error modified pixel values are not associated with a point within a changed 744 region border, then further analysis or testing may be required. For example, a determination can be made as to whether the point associated with the error modified pixel values falls within or external to a region bounded by the warped or changed 218 locations of the reference points (e.g., 318 , 514 - 538 ). For example, such a determination can be made through a coordinate comparison.
- respective distances to respective reference point positions e.g., 318 , 514 , 522 , 524 , 526 , 530 , 534 , 538 ) of the warped marking decision color space (e.g., 510 ) that are associated with a marking decision to place a mark for one and only one colorant, from the position associated with the values of the error modified pixel values can be determined 818 .
- the Euclidian distance from the point associated with the error modified pixel values (c e , m e , y e ) can be calculated to each of the cyan 530 , magenta 534 and yellow reference points of the warped (e.g., 510 ) marking decision color space.
- the region e.g., one of the regions about 640 , 642 or 646 ) associated with the respective reference point position of the warped marking decision color space that is associated with a marking decision to place a mark for one and only one colorant, that is closest to the position associated with the value of the error modified pixel values (c e , m e , y e ) can be identified 822 as the identified 222 point.
- a determination 756 can be made as to whether the error modified pixel values are associated with a point both external to that region and within a region (e.g., 718 , 726 , 728 ) bounded by planes that are tangent to the warped region borders and that extend in a direction perpendicular to the direction of respective axes.
- such a determination 756 can be made by one or more sets of coordinate comparisons. For instance:
- the error modified pixel values must be associated with a point associated with a marking decision to place a mark using one and only one colorant.
- the determination 818 can be made as to which of the warped or changed 218 locations of the reference points is closest to the point associated with the error modified pixel values. Accordingly, while the order in which some of these tests are performed can, to some extent, be selected on the basis of efficiency, which may be a function of likelihood of success and/or computational simplicity, logically, the method may be best understood when the tests are considered in the order outlined in FIGS. 7 and 8 .
- the warping values e.g., 2f i (i 0 )
- the contone value associated with 100% are coverage (e.g., 255) for each of the axes
- all of the reference points overlap and become co-located at 0,0,0.
- a marking decision can be selected 126 under these circumstances by identifying an octant of eight octants originating or radiating from 0,0,0 in which the error modified pixel value is located.
- the modified pixel is located in a W octant and the marking decision associated with the W ( 318 ) octant should be selected 126 . If c e >0, m e ⁇ 0 and y e ⁇ 0, the modified pixel is located in a C octant and the marking decision associated with C ( 330 , 530 ) should be selected 126 . If c e ⁇ 0, m e >0 and y e ⁇ 0, the modified pixel is located in an M octant and the marking decision associated with M ( 334 , 534 ) should be selected 126 .
- the modified pixel is located in a Y octant and the marking decision associated with Y ( 338 , 538 ) should be selected 126 . If c e >0, me>0 and y e ⁇ 0, the modified pixel is located in a B octant and the marking decision associated with B ( 324 , 524 ) should be selected 126 . If c e >0, m e ⁇ 0 and y e >0, the modified pixel is located in a G octant and the marking decision associated with G ( 326 , 526 ) should be selected 126 .
- the modified pixel is located in an R octant and the marking decision associated with R ( 322 , 522 ) should be selected 126 . If c e >0, m e >0 and y e >0, the modified pixel is located in a K octant and the marking decision associated with K ( 314 , 514 ) should be selected 126 .
- This process results in the selection 126 of a marking decision.
- that marking decision is associated with a set of contone values, one for each colorant or separation (e.g., the contone values 0 and 255 are associated with decisions to place no mark and to place a mark respectively).
- Those values are compared to the error modified pixel value being processed to determine a difference.
- That difference or error is combined 122 with received 114 pixel values from one or more yet-to-be processed pixels from an input image and the process is repeated for each pixel in the image to be processed.
- the defined 310 marking decision color space is warped or the location of reference points are changed 218 according to a function of pixel values from the target pixel in the input image and that warped space is used to select a marking decision according to the error modified pixel values associated with that target pixel.
- the method provides the high spatial frequency of mark placement of vector error diffusion while providing the ability to modify edges (enhance or attenuate) which has not been previously provided in vector error diffusion methods. It is noted that these procedures are based on the illustrative marking decision color space definition 310 . Other marking decision color space definitions including, for example other segmentations, might call for other marking decision selection procedures.
- the marking decision color space need only be defined 214 (and optionally 714 ) once.
- that same definition may be used to process multiple images. Accordingly, the processing of a particular image is based on a marking decision color space definition, but does not require that the marking decision color space be defined as part of the processing of that particular image
- an alternative implementation of the algorithm warps the error modified pixel values (c e , m e , y e ) prior to their input to a static, unwarped marking decision color space.
- the warping of input pixel values is in the opposite direction of the warping of the threshold space described above.
- FIG. 9 illustrates the case where the input remains unchanged and the decision space is warped.
- FIG. 10 illustrates the equivalent case where the decision space is left unchanged and the input value is warped.
- a multi-dimensional, vector error diffusion embodiment that takes advantage of this alternative approach can be implemented as follows.
- a marking decision color space (e.g., 310 ) can be defined 214 (and optionally, 714 ) as described above.
- reference points can be located at the extremes of ranges of available colorants.
- region borders can be defined around each of selected reference points.
- region borders can be defined 714 around each reference point that is not associated with a marking decision to place a mark for one and only one colorant.
- spheres can be defined around reference points associated with secondary and tertiary colors (e.g., in the illustrated color space: red, green, blue, white and black).
- a marking decision is selected 126 based on both an error modified pixel value and at least one function of at least one value of the received target pixel contone values.
- warping functions are again defined with regard to axes of the defined 214 marking decision color space (e.g., 310 ). Where the marking decision color space is defined in terms of cyan, magenta, and yellow axes, warping functions can be denoted as f c (•), f m (•), and f y (•). In this alternative embodiment, the warping functions are not used to warp or change 218 , 744 aspects of the marking decision color space, but are instead used to warp or modify values of error modified pixel values.
- the warping functions are used to generate offsets to the error modified pixel values as indicated in the expression below.
- ( c e ′,m e ′,y e ′) ( c e ,m e ,y e )+( f c ( c 0 ), f m ( m 0 ), f y ( y 0 ))
- the modified error modified pixel values can then be tested to determine which region of the marking decision color space the modified values are associated with. For instance, if the marking decision color space was defined 714 to include region borders around selected reference points, tests can be performed to determine if the modified error modified pixel value is associated with a point within any of the defined 714 region borders (e.g., 650 - 658 ).
- the following inequality can be used to test whether the modified error modified pixel values fall within a particular one of the defined 714 region borders: (c e ′ ⁇ c 0 ) 2 +(m e ′ ⁇ m 0 ) 2 +(y e ′ ⁇ y 0 ) 2 ⁇ r 0 2 ), wherein c 0 , m 0 , y 0 identify the location of the selected reference point; c e ′, m e ′, y e ′ are respective values of the modified error modified pixel values; and r 0 2 is the radius of the particular defined 714 region border.
- the modified error modified pixel value is associated with a point within the particular region border and a marking decision can be selected 126 to be the marking decision associated with the reference point and/or associated region border. If the expression is not true for any of the defined 714 region borders, then further testing may be necessary.
- a determination, similar to the determination 756 as to whether the modified error modified pixel values are associated with a point external to a region bounded by the marking decision color space reference points (e.g., 314 , 318 , 322 , 326 , 330 , 334 , 338 ) and within regions bounded by planes that are tangent to the region borders and that are perpendicular to the direction of respective axes (e.g., within 718 , 726 , 728 , etc.) can be made.
- the modified error modified pixel values are associated with a point external to the region bounded by the marking decision color space position of the reference points (e.g., external to the cube defined by the reference points 314 - 338 ) and if the radius of each defined 714 region border is r, then the following tests can be performed to identify a point equivalent to the identification 222 of the point in the warped marking decision space corresponding to the error modified pixel values by instead identifying a warped point in the unwarped marking decision color space.
- the modified error modified pixel value must be associated with a region associated with a primary reference point. That is, in the illustrative example, the modified error modified pixel value must be associated with a point in the marking decision color space 310 that is associated with a marking decision to place a mark using one and only one colorant (e.g., the C 330 , M 334 or Y 338 reference point).
- the modified error modified pixel value must be associated with a point in the marking decision color space 310 that is associated with a marking decision to place a mark using one and only one colorant (e.g., the C 330 , M 334 or Y 338 reference point).
- Which point can be determined by computing, for example, the Euclidean distance between a point associated with the modified error modified pixel values and each of the reference points (e.g., 330 - 338 ).
- the reference point associated with the shortest distance can be identified in a manner equivalent to the identification 222 of a point in the warped marking decision space as corresponding to the modified error modified pixel values.
- the modified error modified pixel values are associated with a point that falls within the region defined by the reference points (e.g., within the cube of the defined marking decision color space 310 )
- distances to each of the primary reference points can be determined in a manner similar to that described with regard to the determination 818 of respective distances to respective reference point positions of the warped marking decision color space.
- the determination would be of distances of a warped position of warped error modified pixel values to unwarped or original positions (e.g., 330 - 338 ) of the primary reference points or reference points associated with a marking decision to place a mark for one and only one colorant.
- the determination as to whether the modified error modified pixel values are associated with a point that falls within or is external to the region bounded by the reference points can be achieved through a comparison of coordinates. For instance, if 0 ⁇ c e ′ ⁇ 255 and 0 ⁇ m e ′ ⁇ 255 and 0 ⁇ y e ′ ⁇ 255, then (c e ′, m e ′, y e ′) is within the region defined by the reference points. It is noted that these procedures are based on the illustrative marking decision color space definition 310 . Other marking decision color space definitions including, for example other segmentations, might call for other marking decision selection procedures.
- the marking decision color space need only be defined 214 (and optionally 714 ) once.
- that same definition may be used to process multiple images. Accordingly, the processing of a particular image is based on a marking decision color space definition, but does not require that the marking decision color space be defined as part of the processing of that particular image.
- the marking decision associated with the identified point can be selected 226 , an error associated with the marking decision can be determined 130 and stored 134 for further processing and the marking decision can be stored or rendered 138 according to the rest of the vector error diffusion process.
- Images can be received 114 via computer communication systems including network connections and storage media.
- image input means can include scanners and/or cameras.
- Computational elements can include computer processors, microprocessors, application specific integrated circuits (ASIC), digital signal processors and other devices.
- Data can be stored before, during and after processing according to the method 110 in computer memory devices and bulk storage devices including, but not limited to, RAM, ROM, EEPROM, FLASH, memory sticks, memory cards, rotating media such as magnetic disks, optical disks and in network based storage devices.
- Functional blocks and/or components can be located in a single device or distributed over a wide geographical area while interacting through communication devices including telephone or computer networks.
- an illustrative system 1110 can include a pixel value combiner 1114 , a marking decision selector 1118 and a color error determiner 1122 .
- the pixel value combiner 1114 receives 114 contone pixel values of a target color pixel from an image to be processed or quantized.
- the pixel value combiner 1114 also receives 118 color contone error values.
- the color contone error values may be received 118 from an error storage device 1126 or directly from the color error determiner 1122 .
- the error may be an initial default error value for use in processing a first pixel, error from the processing of one previously processed pixel or a combination of error or proportions of error from the processing of a plurality of previously processed pixels.
- the received 118 error values are combined 122 with the received 114 target pixel contone values. For example, errors associated with respective color or colorant separations are combined through, for example, addition or subtraction with respective contone values associated with the respective color or colorant separations of the target pixel contone values.
- the marking decision selector 1118 receives the combined 122 values in the form of error modified pixel values. Additionally, the marking decision selector 1118 receives 114 the contone pixel values of the target color pixel. Accordingly, the marking decision selector 118 is able to select 126 a marking decision for the target color pixel based on both the error modified pixel values from the pixel value combiner 1114 and at least one function of at least one value of the received target pixel contone values. The marking decision selector 1118 makes the selection according to vector error diffusion techniques. Accordingly, an error modified pixel value regarding at least a first colorant can be a factor in selecting a marking decision regarding at least the second colorant (e.g., see FIG. 3 or FIG. 5 wherein for a given contone value associated with cyan contone values associated with magenta and/or yellow can change the marking decision selected for cyan).
- the marking decision selector 1118 can define 214 or include a defined 214 marking decision color space (e.g., 310 ) relating marking decisions to contone values regarding a plurality of colorants, the color space being defined, at least in part, by respective reference points (e.g., 314 - 338 ).
- the reference points can be associated with respective extremes of ranges of available contone values and therefore with respective marking decisions having zero error, wherein the marking decision color space associates respective subsets of sets of contone pixel values with marking decisions of respective ones of the respective reference points and wherein the marking decision color space and associated marking decisions being associated with a referenced set of contone pixel values.
- the marking decision selector 1118 can then change 218 the location of at least one reference point in the marking decision color space according to a function of a difference between the referenced set of contone pixel values (e.g., white) and the contone pixel values of the target pixel.
- a location change results in a change in the distance in the marking decision color space between respective subsets of contone pixel values and the relocated reference point or points. Accordingly, the marking decision associated with at least one set of contone pixel values is changed by this relocation 218 .
- the marking decision selector defines 214 or includes a defined 214 marking decision color space wherein region borders are defined 714 around selected reference points.
- the border regions can be defined around reference points that are not associated with a marking decision to place a mark for one and only one colorant.
- region borders 650 - 658 are defined around reference associated with marking decisions to place marks using two, three or zero colorants.
- Region borders encompass points in the marking decision color space to be associated with the marking decision of the respective reference point.
- the marking decision selector 1118 may also change 744 the region border around each respective reference point associated with a region border according to at least one function of a difference between a respective at least one value of the reference set of pixel values and the at least one value of the received contone pixel values of the target pixel.
- region border changing 744 generates respective warped region borders.
- the marking decision selector 1118 can equivalently use a function of the difference between the reference set of contone pixel values and one or more pixels of the target pixel to warp or modify the values of the axes of the marking decision color space 318 or to change, modify or warp the error modified pixel values resulting from the combination 122 of the received target pixel contone values and the error contone values.
- a point is identified in the marking decision color space (e.g., in the warped marking decision color space (e.g., 510 ) or where the error modified pixel value is changed, in the unwarped marking decision color space (e.g., 310 )) that corresponds to the error modified pixel values (or warped or changed error modified pixel values) and a marking decision associated with that point is selected (e.g., 126 , 226 ).
- the marking decision selector 1118 can determine 748 whether the values of the error modified pixel values are associated with a point that falls within the boundary of one of the respective warped region borders. If so, the respective warped region is identified as the identified 222 point and the marking decision associated with the identified region border is selected 226 as the selected 126 marking decision.
- the marking decision selector 1118 may attempt to determine 756 if the error modified pixel values are associated with a point external to a region bounded by the warped or changed 218 marking decision space position of the reference points and within a region bounded by planes that are tangent to the warped region borders and that extend in a direction perpendicular to the direction of respective axes.
- identifying 222 the point in the warped marking decision space corresponding to the error modified pixel values can include identifying 814 the region external to the region bounded by the warped marking decision space position of the reference points and within the region bounded by planes tangent to the warped region borders as the identified point. If such a determination 756 is not made by the marking decision selector 1118 , then respective distances of a point associated with the error modified pixel values to respective reference points can be determined.
- the marking decision selector 1118 can identify 222 the point in the marked warping decision space corresponding to the error modified pixel values by identifying a region associated with the respective reference point of the warped marking decision color space that is closest to the position associated with the error modified pixel values.
- a point in the unwarped or originally defined 214 marking decision color space is identified in a manner similar to that described with regard to identifying 222 the point in the warped marking decision space. However, the determinations or tests are made with regard to the defined 214 / 714 marking decision color space and a warped or modified axes and/or warped or modified error modified pixel values.
- the marking decision selector 1118 may deliver the selected 226 marking decision to a quantized image storage device 1130 and/or to a rendering device 1134 .
- the quantized image storage 1130 may be a computer memory or mass storage device where the marking decision may be accumulated with other marking decisions related to other input image pixels and from where the quantized image may be further processed.
- the rendering device 1134 may be a printer or display device.
- the color error determiner 1122 also receives the selected 226 marking decision from the marking decision selector 1118 .
- the color error determiner 1122 also receives the associated error modified pixel values from the pixel value combiner and compares contone pixel values associated with the marking decisions to the error modified pixel values. Differences between those values are considered error which is either stored in the error storage device 1126 or passed directly to the pixel value combiner to be combined in with values from one or more subsequently processed pixels received 114 from an image data source.
- rendering device or “printer” as used herein encompasses any apparatus such as a digital copier, book making machine, facsimile machine, multi-function machine, display, etc. which performs a quantized print outputting or display function for any purpose.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Processing (AREA)
Abstract
Description
W (e.g., 318): (c l ,m l ,y l)=(0,0,0)
R (e.g., 522): (c l ,m l ,y l)=(0,255−2f m(m 0),255−2f y(y 0))
G (e.g., 526): (c l ,m l ,y l)=(255−2f c(c 0),0,255−2f y(y 0))
B (e.g., 524): (c l ,m l ,y l)=(255−2f c(c 0),255−2f m(m 0),0)
K (e.g., 514): (c l m l ,y l)=(255−2f c(c 0),255−2f m(m 0),255−2f y(y 0)).
W: (c l ,m l ,y l)=(255−2f c(c 0),255−2f m(m 0),255−2f y(y 0))
R: (c l ,m l ,y l)=(255−2f c(c 0),0,0)
G: (c l ,m l ,y l)=(0,255−2f m(m 0),0)
B: (c l ,m l ,y l)=(0,0,255−2f y(y 0))
K: (c l ,m l ,y l)=(0,0,0).
Where (cl,ml,yl) is the translated location of the associated reference point.
WC axis: k c=(255−2f c(c 0))/255
WM axis: k m=(255−2f m(m 0))/255
WY axis: k y=(255−2f y(y 0))/255.
(c−c l)2/(rk c)2+(m−m l)2/(rk m)2+(y−y l)2/(rk y)2=1.
(c e −c l)2/(rk c)2+(m e −m l)2/(rk m)2+(y e −y l)2/(rk y)2≦1
-
- i. if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce≦wc, me≦wm, ye≦wy.
- 2. The following two conditions are true: ne≦wn and (pe−wp)2/(rkp)2+(qe−wq)2/(rkq)2≦1 where n, p and q are one of c, m and y each and n≠p≠q, and where (wc,wm,wy) are the coordinates of the
warped W 318 reference point, identify 814 the regions (e.g., 718) associated withW 318. If neither holds, perform the next test.
- ii. if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce≧tc, me≧tm, ye≧ty,
The following two conditions are true: ne≧tn and (pe−tp)2/(rkp)2+(qe−tq)2/(rkg)2≦1 where n, p and q are one of c, m and y each and n≠p≠q, where (tc,tm,ty) are the coordinates of thewarped K 514 reference point, identify 814 the region (e.g. 728) associated withK 514. If neither holds, perform the next test.
- 1. Exactly two of the following three conditions are true: ce≧tc, me≧tm, ye≧ty,
- iii. if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce≦rc, me≧rm, ye≧ry,
- 2. The following two conditions are true: ce≦rc and (me−rm)2/(rkm)2+(ye−ry)2/(rky)2≦1,
- 3. The following two conditions are true: ne≧tn and (pe−tp)2/(rkp)2+(ce−tc)2/(rkc)2≦1 where n and p are one of m and y each and n≠p,
where (rc,rm,ry) are the coordinates of thewarped R 522 reference point, identify 814 the region associated withR 522. If not, perform the next test.
- iv. if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce≧gc, me≦gm, ye≧gy,
- 2. The following two conditions are true: me≦gm and (ce−gc)2/(rkc)2+(me−gm)2/(rkm)2≦1,
- 3. The following two conditions are true: ne≧gn and (pe−gp)2/(rkp)2+(me−gm)2/(rkm)2≦1 where n and p are one of c and y each and n≠p,
where (gc,gm,gy) are the coordinates of thewarped G 526 reference point, identify 814 the region associated withG 526. If not, perform the next test.
- v. if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce≧bc, me≧bm, ye≦by,
- 2. The following two conditions are true: ye≦by and (ce−bc)2/(rkc)2+(me−bm)2/(rkm)2≦1,
- 3. The following two conditions are true: ne≧bn and (pe−bp)2/(rkp)2+(ye−by)2/(rky)2≦1 where n and p are one of c and m each and n≠p,
where (bc,bm,by) are the coordinates of thewarped B 524 reference point, identify 814 the region the marking decision associated withB 524. If not, perform the next test.
- i. if one of the following holds:
-
- 1. ke<T′=T−f(k0). In this case, the first implementation would select 126 the marking decision associated with the
contone value 0 since ke<T′. The second implementation would select 126 the marking decision associated with thecontone value 0 since ke′=ke+f(k0)<T′+f(k0)=T. - 2. ke>T=T′+f(k0)>T′. In this case, the first implementation would select 126 the marking decision associated with the
contone value 255 since ke>T′. The second implementation would select 126 the marking decision associated with thecontone value 255 since ke′=ke+f(k0)>T+f(k0)>T. - 3. T′<ke<T. In this case, the first implementation would select 126 the marking decision associated with the
contone value 255 since ke>T′. The second implementation would select 126 the marking decision associated with thecontone value 255 since ke′=ke+f(k0)>T′+f(k0)=T.
Sincecases 1, 2 and 3 cover all possible inputs and the selected 126 the marking decision associated with the contone values of both implementations are the same for all three cases, the methods are equivalent. However, warping the input color rather than the marking decision color space (e.g., 310) may reduce computational complexity.
- 1. ke<T′=T−f(k0). In this case, the first implementation would select 126 the marking decision associated with the
W, (c 0 ,m 0 ,y 0)=(0,0,0)
C, (c 0 ,m 0 ,y 0)=(255,0,0)
M, (c 0 ,m 0 ,y 0)=(0,255,0)
Y, (c 0 ,m 0 ,y 0)=(0,0,255)
R, (c 0 ,m 0 ,y 0)=(0,255,255)
G, (c 0 ,m 0 ,y 0)=(255,0,255)
B, (c 0 ,m 0 ,y 0)=(255,255,0)
K, (c 0 ,m 0 ,y 0)=(255,255,255)
(c e ′,m e ′,y e′)=(c e ,m e ,y e)+(f c(c 0),f m(m 0),f y(y 0))
-
- i. if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce′≦0, me′≦0, ye′≦0,
- 2. The following two conditions are true: ne≦0 and pe 2+qe 2≦r0 2 where n, p and q are one of c, m and y each and n≠p≠q,
identify the region (e.g., 718) associated withW 318. If not, perform the next test.
- ii. if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce′≧255, me′≧255, ye′≧255,
- 2. The following two conditions are true: ne≧255 and (pe−255)2+(qe−255)2≦r0 2 where n, p and q are one of c′, m′ and y′ each and n≠p≠q,
- identify the region (e.g., 728) associated with
K 314. If not, perform the next test. - iii. In a manner similar to that described with regard to 814, if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce′≦0, me′≧255, ye′≧255
- 2. The following two conditions are true: ce′≦0 and (me′−255)2+(ye′−255)2≦r0 2,
- 3. The following two conditions are true: ne≧255 and (pe−255)2+(ce−255)2≦r0 2, where n and p are one of m′ and y′ each and n≠p,
identify the region associated withR 322. If not, perform the next test.
- iv. In a manner similar to that described with regard to 814, if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce′≧255, me′≦0, ye′≧255,
- 2. The following two conditions are true: me′≦0 and (ce′−255)2+(ye′−255)2≦r0 2,
- 3. The following two conditions are true: ne≧255 and (pe−255)2+(me−255)2≦r0 2 where n and p are one of c′ and y′ each and n≠p,
identify the region (e.g., 726) associated withG 326. If not, perform the next test.
- v. In a manner similar to that described with regard to 814, if one of the following holds:
- 1. Exactly two of the following three conditions are true: ce′≧255, me′≧255, ye′≦0,
- 2. The following two conditions are true: ye′≦0 and (ce′−255)2+(me′−255)2≦r0 2,
- 3. The following two conditions are true: ne≧255 and (pe−255)2+(ye−255)2≦r0 2 where n and p are one of c′ and m′ each and n≠p,
identify the region associated withB 324. If not, perform the next test.
- i. if one of the following holds:
Claims (17)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US12/577,798 US8351084B2 (en) | 2009-10-13 | 2009-10-13 | Vector error diffusion with edge control |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US12/577,798 US8351084B2 (en) | 2009-10-13 | 2009-10-13 | Vector error diffusion with edge control |
Publications (2)
Publication Number | Publication Date |
---|---|
US20110085218A1 US20110085218A1 (en) | 2011-04-14 |
US8351084B2 true US8351084B2 (en) | 2013-01-08 |
Family
ID=43854634
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/577,798 Expired - Fee Related US8351084B2 (en) | 2009-10-13 | 2009-10-13 | Vector error diffusion with edge control |
Country Status (1)
Country | Link |
---|---|
US (1) | US8351084B2 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100278424A1 (en) * | 2009-04-30 | 2010-11-04 | Peter Warner | Automatically Extending a Boundary for an Image to Fully Divide the Image |
US20170064280A1 (en) * | 2015-09-01 | 2017-03-02 | Samsung Electronics Co., Ltd. | Image processing method and apparatus |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9460495B2 (en) | 2012-04-06 | 2016-10-04 | Microsoft Technology Licensing, Llc | Joint video stabilization and rolling shutter correction on a generic platform |
US9041980B2 (en) * | 2013-01-20 | 2015-05-26 | Qualcomm Incorporated | Subdividing color space for fast vector error diffusion |
JP7277231B2 (en) * | 2019-04-15 | 2023-05-18 | キヤノン株式会社 | Image processing device, image processing method and program |
CN111539972B (en) * | 2020-04-24 | 2023-04-18 | 大连理工大学 | Method for segmenting cerebellar lumbricus in ultrasonic image |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5045952A (en) | 1989-08-21 | 1991-09-03 | Xerox Corporation | Method for edge enhanced error diffusion |
EP0654940A1 (en) * | 1993-11-02 | 1995-05-24 | Xerox Corporation | Method and apparatus for vector error diffusion with output color control |
US7548346B2 (en) | 2000-11-30 | 2009-06-16 | Canon Kabushiki Kaisha | Image processing apparatus and method for performing error diffusion based on density values of different color components |
-
2009
- 2009-10-13 US US12/577,798 patent/US8351084B2/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5045952A (en) | 1989-08-21 | 1991-09-03 | Xerox Corporation | Method for edge enhanced error diffusion |
EP0654940A1 (en) * | 1993-11-02 | 1995-05-24 | Xerox Corporation | Method and apparatus for vector error diffusion with output color control |
US5621546A (en) * | 1993-11-02 | 1997-04-15 | Xerox Corporation | Method and apparatus for vector error diffusion with output color control |
US7548346B2 (en) | 2000-11-30 | 2009-06-16 | Canon Kabushiki Kaisha | Image processing apparatus and method for performing error diffusion based on density values of different color components |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100278424A1 (en) * | 2009-04-30 | 2010-11-04 | Peter Warner | Automatically Extending a Boundary for an Image to Fully Divide the Image |
US8885977B2 (en) * | 2009-04-30 | 2014-11-11 | Apple Inc. | Automatically extending a boundary for an image to fully divide the image |
US20170064280A1 (en) * | 2015-09-01 | 2017-03-02 | Samsung Electronics Co., Ltd. | Image processing method and apparatus |
US10063832B2 (en) * | 2015-09-01 | 2018-08-28 | Samsung Electronics Co., Ltd. | Image processing method and apparatus |
Also Published As
Publication number | Publication date |
---|---|
US20110085218A1 (en) | 2011-04-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7382915B2 (en) | Color to grayscale conversion method and apparatus | |
US8351084B2 (en) | Vector error diffusion with edge control | |
US8482804B2 (en) | Image processing apparatus and image processing method | |
JPH0380767A (en) | Gradation recorder for image | |
JP4170353B2 (en) | Image processing method, image processing apparatus, image reading apparatus, image forming apparatus, program, and recording medium | |
JPH05501341A (en) | Color digital halftone gradation processing method using vector error diffusion method | |
US20060232798A1 (en) | Blended error diffusion and adaptive quantization | |
JP2008017051A (en) | Image processor, image processing method, and program | |
US20070081205A1 (en) | Image recording apparatus and method providing personalized color enhancement | |
US8995021B2 (en) | Black trapping methods, apparatus and systems for binary images | |
US6972869B2 (en) | Mapping color to colorant amounts when colorants have similar hue | |
JP2015173419A (en) | Image processing device, control method of the same and program | |
JP2006115500A (en) | High-speed low memory paper color suppression algorithm | |
US10200568B2 (en) | Pixel processing | |
EP1492328A2 (en) | Method for color rendering using scanned halftone classification | |
KR20060051730A (en) | Data embedding scheme for duplex color laser printer | |
Park et al. | Halftone blending between smooth and detail screens to improve print quality with electrophotographic printers | |
JP6474315B2 (en) | Image processing apparatus, image forming apparatus, image processing method, image processing program, and recording medium therefor | |
US7502145B2 (en) | Systems and methods for improved line edge quality | |
JP2006155572A (en) | Image processor, image forming device, image processing method and computer program | |
JP4504327B2 (en) | Edge detection for distributed dot binary halftone images | |
KR20070063411A (en) | Noise reduction for primary tones for image replication systems | |
US20060221397A1 (en) | Method and apparatus for forming image using edge information of image | |
JP4259410B2 (en) | Image processing device | |
US10425558B2 (en) | Methods and systems for processing documents using multiple halftone screening techniques |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: XEROX CORPORATION, CONNECTICUT Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:BERNAL, EDGAR A.;LOCE, ROBERT P.;REEL/FRAME:023360/0666 Effective date: 20091012 |
|
FEPP | Fee payment procedure |
Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
FPAY | Fee payment |
Year of fee payment: 4 |
|
FEPP | Fee payment procedure |
Free format text: MAINTENANCE FEE REMINDER MAILED (ORIGINAL EVENT CODE: REM.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
LAPS | Lapse for failure to pay maintenance fees |
Free format text: PATENT EXPIRED FOR FAILURE TO PAY MAINTENANCE FEES (ORIGINAL EVENT CODE: EXP.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
STCH | Information on status: patent discontinuation |
Free format text: PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362 |
|
FP | Lapsed due to failure to pay maintenance fee |
Effective date: 20210108 |